JPH0218947A - Holder and compensation system of non-repeated error - Google Patents

Abstract

PURPOSE: To improve the accuracy of measurement data by compensating for the error caused by a measurement data holder. CONSTITUTION: A movable holder 12 being fixed to a sample 14 detachably therefrom is fixed with a reference 28 for movement. A first sensor 16 generates a measurement signal when the sample is shifted by the movable holder and a second sensor 26 generates an error signal when the reference 28 is shifted by the movable holder. More specifically, for each point of the sample at which the first sensor 16 generates a measurement, the second sensor 26 generates an error signal from a corresponding point of the reference. The measurement signal is combined with the error signal and the error of the measurement signal is compensated thus producing a highly accurate output signal. According to the arrangement, mechanical errors of the holder can be compensated for effectively and the measurement accuracy of an object is enhanced.

Description

[Detailed Description of the Invention] <Industrial Application Field> This invention relates to the field of measurement and testing,
In particular, the present invention relates to a system for compensating for errors mechanically introduced into measurement data by a device that moves a measurement object from which data is obtained.

<Conventional technology> Flatness and looseness (bow) of semiconductor wafers,
The warp or other characteristic must be strictly within predetermined criteria for the wafer to be usable.1 Determining whether, and to what extent, an individual wafer conforms to the criteria. to be measured. but,
At present, measurement errors occur due to mechanical errors in the holder that holds the wafer during the measurement, and as a result, there is a limit to the accuracy of wafer measurement.

Therefore, a method can be considered that uses a wafer test device manufactured with high precision so that virtually no errors occur.

However, such high precision devices often have limitations in processing efficiency and disadvantages such as high costs associated with installing and maintaining the devices.

- There is a method of electronically processing the force measurement data to compensate for device-induced errors in the measurement data as long as the errors are repeated. US patent application Ser. In the rose/work station embodiment, the x, θ, bidirectional movable device is calibrated to determine the error due to its X and θ position repeated from measurement cycle to measurement cycle. - once the instrument is calibrated, the measurement data is compensated;
It exhibits accuracy that is not limited by repetitive x, theta errors associated with the mechanical nature of the device.

<Summary of the Invention> The present invention aims to provide a measurement system that separates non-repeatable errors caused by a device from measurement data, compensates for non-repeatable errors caused by the device in measurement data, and thereby improves the accuracy of measurement data. It is. The device according to the invention for receiving the specimen to be measured has a movable holding device on which the specimen is removably mounted. The movable holding device is fitted with a support for movement. The first sensor generates a measurement signal when the specimen is moved by the movable holding device. A second sensor then generates an error signal when the reference is moved by the movable holding device. Thus, for each point on the specimen at which the first sensor outputs a measurement signal, the second sensor outputs an error signal from the corresponding point on the reference.

The measurement signal and the error signal are combined to compensate for non-repetitive errors in the measurement signal and provide a highly accurate output signal. In a preferred embodiment, the output signal is digitized, non-repeat error is compensated for, and the desired object property or properties are output with very high accuracy. When combined, complete error compensation is achieved.

The purpose, summary, and advantages of the present invention are as follows:
This becomes clear from the description and drawings of non-limiting preferred embodiments.

<Example> The present invention is effective in the following applications.

That is, a specimen to be measured is removably held in a holding device, which in cooperation with a sensor moves the specimen and the sensor relative to each other to obtain a measurement result of the specimen in the form of an output signal from the sensor. determine one or more properties of the specimen from the results, and the results are subject to errors due to imperfections in the mechanical position of the holding device, which errors contaminate the output of the sensor;
This makes the invention particularly useful in fields where there is a limit to the accuracy of determining desired properties. This is useful in situations where the sensor element is subject to momentary natural deviations in movement from its position, which introduce errors into the data measured by the sensor element. The invention therefore provides means for providing a reference representative of the ideal position of the device, means for measuring deviations of the device from the reference position, and means for compensating for measured data for deviations from the ideal position. It is intended that In a preferred embodiment, the present invention provides x, θ,
The disclosure relates to a vacuum chuck movable in the Z direction and a cooperative capacitive sensor element for measuring the proximity of a specimen to a probe. As will be appreciated by those skilled in the art, the measured data yields characteristics such as wafer flatness, unevenness/warp, etc., and the present invention, which compensates for equipment-related errors in the z11 constant data, provides such characteristics with very high precision. make it possible to provide. Other applications and configurations are also contemplated by the invention, which is not limited to merely illustrative and preferred embodiments.

FIG. 1 is a block diagram of a preferred embodiment of the apparatus and non-repetitive error compensation system of the present invention.

System 10 includes a vacuum chuck 12 for releasably holding a wafer 14 (or other specimen), and also includes a sensor 16 located proximate to vacuum chuck 12. Sensor 16 is operative to measure the distance to a predetermined surface area of wafer 14 that is brought into close proximity to sensor 16 .

The wafer 14 has characteristics such as flatness and loose/warp profile determined by measurements of points located on the sensor 16. Sensor 16 is preferably I
tA" and the first probe 18 indicated by "A".
B", these probes constitute a pair of measurement heads 22. The first probe 18. The second probe 20 is preferably a capacitive sensor, the first probe 18 is a support. The second probe 20 is similarly fixed to the support body 24 .

The first probe 18 and the second probe 20 may be movable relative to each other, so that the size of the measurement head 22 can be adjusted.

The reference probe 26 is attached to a support 24 on which the first probe 18 and the second probe 20 are attached.

In this way, all probes are fixed to each other.

Any means of mounting the probe in this manner is applicable.

Any type of probe can be used, but the U.S. Patent No. 3.99, which describes capacitive thickness measurement of non-grounded elements,
Preferably, the capacitive measurement system disclosed in No. 0.005 is employed.

A reference 28 is attached to the vacuum chuck 12 and is configured to move as the vacuum chuck 12 moves in one or more degrees of freedom. In a preferred embodiment, the reference 28 may be a metal disk and is mounted to move with the vacuum chuck 12 in the x and θ directions.

In other embodiments, reference 28 may simply be a bar moving in the X direction.

Reference probe 26 defines the ideal position of vacuum chuck 12 via reference 28 . The output of the measurement by the reference probe 26 indicates the distance from the ideal position.

Analog signal conditioning and synthesis electronics 30 are illustrated by dashed blocks. This electronic circuit 30 is connected to the outputs of the first probe 18, the second probe 2o, and the reference probe 26.

The electronic circuit 30 connects the first probe 18. Second probe 20
, any configuration that adjusts the analog signals from the reference probe 26 according to the characteristics of the wafer to be measured and synthesizes them in a manner that compensates for non-repetitive mechanical imperfections in the equipment. The signal combination of the first probe 18 and the second probe 20 may be selected to conform to the desired characteristics of the wafer being measured, as will be readily understood by those skilled in the art. For example, if they are added, the thickness of the wafer is obtained, and if they are subtracted, the distance to the center line of the wafer is obtained. In order to obtain a highly accurate rose/warp profile, the synthesis and adjustment circuit may include, for example, linearizers 32, 34, and 36 connected to the first probe 18, the second probe 2o, and the reference probe 26, respectively. The difference between the outputs of the first probe 18 and the second probe 20 is calculated in an analog adder 38. The output from the reference probe 26 is combined with the output of the analog summer 38 to compensate for non-repetitive device errors in the measured data in real time, thereby providing a compensated output signal on the wafer 14 as shown in FIG. relative to the position of the lower reference 28 and reference probe 26.

A mixing element 40, such as an analog adder, is provided to convert the output of the reference probe 26 to "REF".
Subtracted from the reference level indicated by , this reference level represents the nominal position of the device and the deviation from it is
2 indicates a mechanical deviation from the ideal position specification. The compensated output signal is sent to the A/D converter 42.
is input.

The Xθ2 assembly 44 is installed to rotate the vacuum chuck 12 two θ ramps around the axis and to move the vacuum chuck 12 along the X axis and along two axes. The Xθ2 assembly 44 responds to a plurality of control signals from the processor 46 to control the vacuum chuck 12 to move in the x and θ directions to successively bring predetermined points of the wafer 14 closer to the measurement head 22. position. This point is preferably selected to cover the entire wafer 14, for example for a rose/warb profile. Although any configuration can be used for the Xθ2 assembly 44, the
No. 572.6, entitled Wafer Flatness Station, is preferred, and any movement pattern for operating the Xtheta assembly 44 may be employed.
No. 95 and U.S. Pat.
An exemplary operating sequence can be obtained from No. 49. Processor 46 may be any suitable processor, for example included in the device and connected to A/D converter 42 via data/address bus 48. Processor 46 connects conventional latch driver 50 via data/address bus 48.
connected via. Processor 46 includes RAM 52 and PROM 54, which are coupled in conventional manner. Central control processor 56 connects data/address bus 48 to a communication link, preferably IE.
Preferably, it is connected via an EE488 bus 58 and an IEEE488 interface 60. processor 46
is preferably a slave processor of the central control processor 56, but may of course be a separate processor.

Next, the operation will be explained.

The Xθ2 assembly 44 controls the vacuum chuck 12 to move the wafer 14 held by the vacuum chuck 12 to position each different point on the wafer 14 within the measurement head 22 for measurement, at each point. 1st
The probe 18 and the second probe 20 output signals. This signal indicates the distance to the opposite corresponding surface of wafer 14 at each point. Reference probe 26 operates simultaneously and outputs a signal. This signal indicates the distance to the surface of the corresponding reference 28 at each point on the wafer 14 described above. If vacuum chuck 12
If the reference probe is in its correct position, the reference probe 2
The signal from 6 corresponds to the nominal position signal. However, if the vacuum chuck 12 is subject to non-repeatable errors and is not in its original position in two directions, the reference probe 26
The output of is changed accordingly, and this change is provided by mixing element 40 as an error component output signal. Electronic circuit 30 subtracts this error component signal from the measurement signal in real time as the signal is output, and the compensated signal is then digitized by A/D converter 42. As will be appreciated by those skilled in the art, the data thus obtained can be processed by a processor to obtain one or more characteristics of the wafer, such as flatness, bulge/warp, etc. Reference may be made to the processing sequence and algorithm for obtaining the flatness profile of a wafer from Application No. 572.695 entitled Wafer Flatness Station. Also disclosed is an apparatus and method for separating mechanical errors from a measured characteristic and compensating for errors in the measured characteristic, and a similar application tube 802 called a rose/warp station.
No. 049, reference may be made to an algorithm that provides a wafer rose/warp profile that is free from processing sequence and repetition errors. Other sequences and algorithms for determining flatness para/warp and other wafer characteristics can be readily implemented by those skilled in the art.

The present invention can eliminate errors caused by non-repetitive deviations from ideal position specifications of the device that contaminate the measurement signal from the probe. Additionally, errors caused by repeated deviations from the ideal positional specifications of the device can be eliminated by the techniques disclosed in the above-mentioned application Ser. No. 802,049. In this manner, complete error compensation for both repetitive and non-repetitive displacements of the Xθ2 assembly 44 in one or more degrees of freedom can be achieved.

This dramatically improves data accuracy despite mechanical imperfections in the equipment and non-ideal conditions associated with out-of-tolerance limits.

Various modifications of the invention will be apparent to those skilled in the art without departing from the scope of the claims.

Next, an embodiment of the apparatus for compensating for the above-mentioned repetition error will be described.

First, an overview of the warp measuring device will be explained.

As shown in FIG. 2, a warp measurement device para/warp measurement device 216 forms part of system 210. The system 210 includes a central processing unit 212 .

The central processing bag W212 is connected to the alignment device 214. It is connected to and controls the rose/warp measuring device 216 and sample mover 218. The sample mover 218 is, for example, a pair of parallel bar belts, etc., and moves a sample such as a semiconductor wafer to the alignment device 214, the bow/warp measurement device m! 216. This belt may be, for example, as shown in U.S. Patent Application No. 725,159;
The sample is placed on the sample mover 218 at the top left of one drawing. The elevator is, for example, U.S. Patent Application No. 37
The type shown in No. 9559 may be used.

Central processing unit 212 controls sample mover 218 to transport each wafer to alignment device 214 . Although various other devices can be used to align the wafer, it is preferable to use the device shown in U.S. Pat. No. 4,457,664. Perform centering and orientation. The central processing unit 212 then sends a signal to the sample mover 218 to
The wafer is transferred to the wafer receiving device of the warp measurement device 216. The uneven/warb measuring device 216 then measures the profile (warpage) of the wafer.

FIG. 3 shows details of the rose/warb measuring device 216. The rose/warb measuring device 216 is connected to a vacuum chuck 222.
It is equipped with The vacuum chuck 222 is the wafer 2
24 (or a sample which is an object to be measured) is a support device that removably holds the object to be measured. The vacuum chuck 222 is configured so that its relative position to the wafer 224 can be changed. The sensor 226 is located near the vacuum chuck 222 and is configured to be able to measure the distance to the surface of the wafer 224 within an arbitrary range.

Naturally, the warping of the wafer 224 is unknown. The sensor 226 has a first probe 228 and a second probe 23
0, between which a capacitive measuring head 232 is set. It is desirable that the first probe 228 and the second probe 230 are capacitive sensors.

A second probe 230 is secured to a support 234. First probe 228 is attached to arm 236. The arm 236 is attached to the support member 238 via a screw, and is movable with respect to the support 234, so that the interval between the capacitive measuring heads 232 can be adjusted. An analog signal conditioning unit 239 is connected to the first probe 228 , the second probe 230 , and the first probe 228 . The sum of the outputs of the second probe 230 A+
It is designed to output an analog signal representing B. That is, the first probes 228 and 229 output measured values at a plurality of predetermined points on the wafer sequentially located on the capacitive measuring head 232, and the analog signal conditioning unit 23
9 outputs the sum of the measured values of the first probe 228 and the second probe 230 at each measurement point. The analog signal adjustment unit 240 is similarly connected to the first probe 228 and the second probe 230, and outputs the difference A-B between the measured values of the first probe 228 and the second probe 230 at multiple points on the wafer. . A+B represents the thickness of the wafer, and A-B represents, for example, the distance to the center line in the thickness direction of the wafer. The multiplexer 241 is connected to the analog signal conditioning unit 239. It is connected to the output side of the analog signal conditioning unit 240. The A+B output is useful for flatness verification, as described in US Patent Application No. 572,695.

On the other hand, the AB output is used for warpage measurement and its compensation.

An A/D converter 42 is connected to the multiplexer 241 and digitizes the signal from the multiplexer 241. Although various other probes can be used as the first probe 228 and the second probe 230, US Pat.
Preferably, a capacitive measuring system as described in US Pat. No. 990,005 is used.

x, θ, 2 assembly 44 is vacuum chuck 22
2, and the vacuum chuck 222 is rotated twice through the θ ramp around its center. Also, mXp θ that moves the vacuum chuck 222 in the X-axis direction and the Z-axis direction
, the Z assembly 44 operates the vacuum chuck 222 in response to the control signal to remove the capacitive measurement head 2.
A plurality of predetermined positions of the wafer 224 are performed between the positions of the wafers 224 and 32. This x, θ preferably covers the entire area of the wafer 224 in order to collect the first to fourth databases to be described later.

Although various configurations can be used as the Z assembly 44, it is preferable to use the one described in the above-mentioned US Pat. No. 4,457,664.

A processor 46 is connected to the A/D converter 42 via a data bus 48 and is connected to an x, theta, Z assembly 44 via a data bus 48 and a conventional driver 50. puot=sa 46 is R
AM52. It has PROM54. Central processing unit 212 preferably interfaces with IEEE 488 bus 58 and IE
It is connected to the data bus 48 via an EE488 interface 60 bus system. Processor 46 is responsive to central processing unit 212 (i.e., processor 46
is a slave, and the central processing unit 212 is a mask.

), position a predetermined point near the center of the wafer between capacitive measurement heads 232, and then position a point around the center of the wafer. At the same time, the processor 46 reads the output at each point of the A/D converter 42 and writes it to the RAM 52 via the data bus 48.
This corresponds to the spatial position of the measurement point on the wafer set by .

Processor 46 outputs control signals for x, θ, Z assembly 44 to driver 50 via data bus 48 in response to commands from central processing unit 212 . This controls the X step motor, θ step motor, Z actuator, and vacuum conditions. In response to the X control signal, the shaft of the X step motor rotates, causing the worm gear to rotate. The screw of the worm gear is screwed into the screw of the housing which is slidably mounted on the linear guide rail. This allows vacuum chuck 2
22 is adjusted in its position along the X-axis. Similarly, θ
The directional position is also adjusted by the theta step motor and belts and gears in response to the theta control signal. Further, the position of the vacuum chuck 222 in the Z direction is also adjusted by two actuators in response to the Z signal. vacuum chuck 22
The ON/OFF state of the vacuum supplied to 2 is controlled by the vacuum line. When ON, the wafer is attracted by the vacuum chuck 222, and when OFF
At this time, the wafer is released from the vacuum chuck 222.

Next, the operation of the processor 46 will be explained based on FIG. 4. Instructions from the central processing unit 212 are received by the processor 46 (62).

Processor 46 decodes this command and returns x, θ.

Z and P code to drive vacuum actuator
Fetch from ROM54. As a result, the wafer is sequentially moved to a plurality of predetermined points within the capacitive measuring head 232 (66), and the processor 46 continues to operate until the execution of this instruction is completed (68).
is an analog signal adjustment unit 240 at each measurement point on the wafer.
The measured value from the corresponding first one in the A/D converter 42
to one of the fourth tables.

In the table, the data is stored at a predetermined address corresponding to the position of each point on the wafer determined by the rotational position of the X motor and the θ motor (70).After measurement, the processor 46 transfers the data to the central processing unit 212. (72).

Here, mechanical errors that have conventionally occurred in measuring devices will be explained with reference to FIG.

The drive in the X-axis direction has a rail guide carriage 76. This rail guide carriage 76 moves according to a screw of an X worm (not shown) connected to the shaft of an The rail is corrugated as shown by relief line 80. The variation due to this becomes maximum as shown by arrow 82. In actual equipment, this error is on the order of ±0.004 inches (±101.6 microns). Such variations displace the position of wafer 224 in the Z direction, resulting in an AB error component. This is directly proportional to the difference from the ideal rail.

The above is an error in the X-axis direction, but the same is true for the θ direction.
The θ drive device has a θ step motor connected to the vacuum chuck 222 via a belt or gears, and is thereby configured to rotate in the θ direction about the Z axis. An ideal axis of rotation 84 and an actual axis of rotation 86, with the ideal axis of rotation indicated at 84 and the actual axis of rotation indicated at 86.
The difference in x appears as an error in two directions of the wafer 224,
.

It changes as the θ direction changes. Such variations are caused by the mechanical precision of the bearing, the precision of the perpendicularity of the wafer and the shaft, and the like. In the actual device, 150mm
Taking a wafer as an example, the two-way error in the vacuum chuck 222 is approximately ±0.0022 inch (±56
．． 25 microns).

The present invention aims to eliminate such mechanical errors in the measuring device, and makes it possible to non-mechanically separate the error component in the X direction and the error component in the θ direction. this is,
This can be achieved by driving the x, θ, and Z assembly 44 multiple times, and the error component in the X direction can be specified by fixing the θ coordinate, and the error component in the θ direction can be specified by fixing the X coordinate. .

x-calibration> First, a method for removing errors in the X direction will be explained.

Eliminating errors in the X-axis direction is most easily achieved by measuring the same point on the wafer twice. That is, by measuring the error component twice with opposite polarities, the error component can be easily calculated from the difference between the measured values.

The wafer is sequentially moved by the x, θ, Z assembly 44 and measured multiple times in a predetermined direction from the reference point of the wafer. Here, the measured value M, (x, θ) at each position on the wafer obtained by the sensor is is the component Mw(X.

θ) and a component Mc(X+θ) related to unnecessary equipment, and is expressed by the following formula.

M, (x, θ) = Mv (x, θ) + Mc (x, θ) --
--(1) Here M, (x, θ), My(x, θ
), Me (X.

θ) is a matrix.

This M□(x, θ) data 88 shown in FIG. 6a is stored in the RAM.
52 is temporarily stored. At this time, each measurement point on the wafer is associated with a specific address. Each measurement point is x, θ, and X of the Z assembly 44 based on a predetermined home position.

It is defined by x and θ coordinates indicating the θ position. x, θ, Z assembly 44 in sequence as shown by arrow 90 in FIG. 6a.
M, (X) for multiple wafer positions located between sensor heads by.

θ) Data is serially stored in the RAM starting from address (0, 0) until the data table is full.

Once all positions have been measured, the wafer is flipped over and measured again from its home position.

Each measurement point on the wafer is positioned on a capacitive sensor head, and measurements of M, (x, θ) are performed.

This Ml(x, θ) is Mw(X) related to the wafer.

θ) and components of Mc(x, θ) related to the device itself. Since this measured value was measured at the same point from the back side of the wafer, the polarity of Mv(x, θ) is opposite to that obtained by measuring the front surface. This is shown in the formula below.

Mz (x, θ) = Mw (x, - θ) + Mc (X - θ
) - (2) Here, Mz (x, θ), My (x, -〇)
, and M c (x, θ) are matrices.

As shown in FIG. 6b, the M2 data is temporarily stored in the RAM 52. That is, each point on the wafer is sequentially positioned on the capacitive sensor head, and the measured values are sequentially written to predetermined locations as shown by arrow 94 in the opposite manner. Each measurement point is based on the home position as in the above case x
, θ, defined by the (x, θ) coordinates of the Z assembly 44.

By adding the matrices of Ml and Ml data, M c (X, θ) is separated as shown in the following equation.

This Me(x, θ) is an error component related to the X direction of the chuck.

M, + M, = M w (X, θ) − My
(x, -〇)+Mc(Xsθ)+Mc(x,θ)...
・・・・・・・・・・・・・・・(3) Rearranging this, Me(X, e)=(M−(X, θ)”Ml
(X −8))/2−− (4).

θ-Calibration> Measure the flipped wafer for 1M, and when the data table is filled, the wafer is released from the chuck.

Only the chuck rotates in the θ direction by a predetermined step θ. This position is set as the second home position, which is shifted by θ from the original first home position.The wafer is turned over to the next position, and the wafer is again attracted to the chuck. The measurement is taken again. This measurement value is shown as follows.

M3 (x, θ)=Mw(X, e)+Mc(x,
8-e xE"" (5) As shown in FIG. 6C, the M data is temporarily stored in the RAM 52. The value at each measurement point on the wafer is stored at each corresponding address in the RAM.

Each address is based on the offset home position x
, θ, Z are defined by the X-θ coordinates of the assembly.

When the measurement of the flipped wafer based on the offset home position is completed, the wafer is released from the chuck again, and only the chuck rotates by a predetermined number of steps (θ) from 01. This position is set as the third home position. The wafer is again attracted to the chuck, the same point on the wafer is positioned between the capacitive sensor heads, and a measurement is made based on the third home position.The measurement result is expressed by the following equation.

M4(x, θ)=Mw(x, θ)+Mc(x, θ−θ,
) (6) M4 data is stored in the RAM 52 as shown in FIG. 6d. Each measurement value is x, θ, and X of the Z assembly based on the third home position.

It is associated with a specific address indicating the θ position. Calculations for separating the Z error component due to θ from the data obtained above are shown below.

M2(X, θ)−My(x, θ)=Mv(x, θ)+
Mc(x, θ)-Mw(X, θ)-Me(x,
θ-θ, )・(7) Putting this in order, iyo(θ)=Mc(θ)-Me(θ-θ,)...
It can be expressed as (8).

Here, iyo(θ) is a variable equal to the left term of equation (7).

The Fourier expansion of (8) yields t(ω): W(ω)-W(ω-θ,)...
... It is expressed as (9).

Here, the work t(ω) is i (θ), and W(ω) is Me(θ).
), W(ω-01) is the Fourier expansion of M c (θ-θ, ).

Equation (9) can be written as follows.

Next, subtracting equations (2) and (6), M2 (x, θ) - M4 (x, θ) = Mv (x, θ) + M
c(x, θ)-My(x, θ)-Mc(x, θ-02)
......(11) is obtained.

Here, L(ω) is W(ω) and W(+,
l-02) are respectively Mc (θ) and Mc(θ-θ2
) is the Fourier expansion of

Equation (13) can be written as follows.

Equations (to) and (14) can be summed by using pre-selected weighting functions to remove unnecessary noise effects. The function is preferably in the form of a linear combination as shown below.

I3(ω)=α(ω)■,(ω)+β(ω)■,(ω)
......(15), rearranging this (12)
Expressed by the formula.

1yo(θ)=Mc(θ) −Me(θ−02)−・・・
...(12) Here, it(θ) is an intermediate variable equal to the left term of (11).

Fourier expansion of equation (12): L(ω)=W(ω)-W(ω-02)...
(13) Here, I and (ω) are the third variables and are equal to a linear combination.

Also, α(ω)=1δ12γ/(1γ12+1612)
β(ω)=1γ12δ/1γ12+1δ.

Inverse Fourier expansion W of equation (15), (x, θ) is x, θ.

It shows the θ-dependent error component of the Z assembly.

The step on of the theta motor is preferably chosen equal to 100 steps, and it is important that the offsets in the theta coordinate, ie, θ□, θ2, are chosen to be disjoint with respect to each other and on. Otherwise, δ and γ will not hold.
In the preferred embodiment, θ1 is 17 steps and θ is 23 steps.
selected for the step.

Next, the operation of the entire apparatus will be explained based on the flowchart of FIG. A flowchart 104 shows the operation of the central processing unit 212. The central processing unit first sends a command to the alignment device 214 to adjust the position and orientation of the wafer (106), and then continues until the alignment is completed. Wait (108), and then the central processing unit moves the aligned sample to the sample mover 218, and moves the center of the sample to the capacitive sensor head 23.
2 (110), wait until the operation is completed (
112), the central processing unit then sends commands to the processor of the rose/warp measuring device.

A measurement of the center of the sample is performed (114).

The central processing unit waits until it receives the measured value (116).
), the sample center measurement command is executed by the rose/warb measuring device's processor based on instructions stored in its FROM control table.

That is, as shown in FIG. 8, the processor of the rose/warp measurement system moves the chuck from its home position along the
At this point, the first processor that attracts the wafer operates the X step motor to position the center of the wafer, indicated by a point 120, between the capacitive sensor heads. Then, the processor releases the wafer and commands the x and θ step motors to return to the original home position.

Then, the processor attracts the wafer and sequentially positions three points 122, 123, and 124 located on the inner circumference of the wafer between the sensor heads. The measured values of each point 122, 123, 124 are temporarily stored in RAM.

Returning to FIG. 7, the central processing unit stores the wafer data sent from the loose/warp measurement device in each position of the M1 data table (FIG. 6a) (126), and the central processing unit stores the wafer data sent from the loose/warp measurement device in the sample mover 218. Sending commands to align the center of the sample with the center of the chuck at its home position (128) and waiting for the end of the operation (130), the central processing unit then sends commands to the processor of the rose/warp measurement device. Measurement is performed on the periphery of the center of the wafer (132).

In response to this command, the processor fetches x, θ, Z and vacuum signal codes for peripheral measurement from FROM. The processor then sends control signals to the x, θ, 2 and vacuum drives to sequentially position predetermined measurement points on the periphery of the center of the wafer between the capacitive measurement heads. The processor then executes the "periphery measurement" command as shown in the embodiment of FIG. The area surrounding the center of the sample is measured at predetermined points in a ring shape sequentially from the outer circumference toward the inner circumference.

Preferably, the central processing unit instructs the processor of the rose/warb measuring device to measure each ring in turn, the processor driving the vacuum.

The wafer is attracted at its center point, the X step motor is driven, the chuck is moved so that the outermost ring 136 is located between the sensor heads, and the processor
The step motor is driven to position a plurality of predetermined points on the ring 136 (three points are represented by dots in FIG. 8) close to each other between the capacitive sensor heads. The processor rotates the theta step motor so that measurements are taken in a complete ring. At each measurement point, the processor converts the reading of the A/D converter 42 into R
Store in the AM52 data table.

This storage address is predetermined and corresponds to a point on the ring 136 defined by the rotational positions of the θ motor and the X motor with respect to the home position. The number of steps of the θ motor is preferably 1°O per rotation.

Once the measured values for each measurement point on the outermost ring have been stored in the local memory of the rose/warp device, the processor transfers this data to the central processing unit (72 in FIG. 4).

Returning to FIG. 7, the central processing unit waits for the data of the outermost ring to be accumulated (138).

Store data representing the distance from the center line of the wafer at each point (140). When data collection for the innermost ring is not completed (142), the central processing unit instructs the processor to continue rotating the wafer. Send a command to move to the next inner ring 144 (144)
, the processor receives this instruction and executes it. After the data on the innermost ring has been accumulated, the central processing unit sends a command to the processor to stop the rotation of the sample, return the vacuum chuck to its original position, release the sample, and wait for the operation to complete (148).

Next, the central processing unit stores the data transferred from the rose/warp measuring device in the memory of the first data table M1 (150).

Next, the wafer is turned over and returned to the sample mover, and the steps 106-116, 126-132 described above are followed.
138 to 150 are repeated by the central processing unit (
152), the data of the flipped wafer is the reverse of the data on the front side, and is stored in the second data table M2 in the RAM of the central processing unit in a similar manner.

The central processing unit then sends a "01 rotation" command to the processor (154) and waits for this command to be executed (
156).

Once the .theta. rotation has been completed, preferably 17 steps, the procedure for data acquisition of M, M, described above is repeated. These data representing the distance to the center line of the wafer at a position shifted by 01 from the home position are stored in the R of the central processing unit.
Stored in the third data table M3 in AM (15
8).

Furthermore, the central processing unit performs the same operation as above.
The data is stored in the fourth data table M4 in AM (160, 162, 164). At this time, the deviation from the home position of the 8 drive units is preferably 23 steps.

The central processing unit then uses the measured components of the wafer itself to
Separate directional errors (166).

This is done by executing the above equations (1) to (4) using software, and the error separation in the θ direction is performed using the above equations (5) to (1).
5) (168), these errors are stored in memory.

The central processing unit removes the x and θ error components.

Only the measured values of the wafer itself are left (170), and this data is processed to obtain the wafer rose/warb values (17).
2) The warp calculation is typically obtained by executing the following formula on a computer.

Warp (x, θ) = M1 (x, θ) - (Me (x, θ)
10 W, (x, θ)) FIG. 9 is a flowchart showing the sequence of rose/warp measurement. The wafers sequentially sent to the rose/warp measuring device are first measured for M, (x, θ) data, and the X calibration value M c (X, θ) already obtained in the above procedure is used.
) and θ calibration value W, (x, θ) (184
), the compensated data represents the rose/warp profile values with high accuracy.

The present invention can be modified in various ways other than those described above by those skilled in the art.

<Effects of the Invention> As explained above, according to the present invention, mechanical non-repetitive errors of the device can be effectively compensated. Therefore, there is an effect that the measurement accuracy of the target object can be significantly improved.

[Brief explanation of drawings]

Fig. 1 is a drawing showing a preferred embodiment of the measuring device and compensation system of the present invention, Fig. 2 is a block diagram showing an example of a rose/warp device, and Fig. 3 is a drawing showing a preferred embodiment of the measuring device and compensation system of the present invention. A block diagram showing one embodiment; FIG. 4 is a flowchart showing the operation of the processor of the rose/warb device; FIG. 5 is a schematic diagram of the measuring device illustrating the occurrence of mechanical errors; FIGS. 6a to 6d. The figure is a schematic diagram showing a memory diagram corresponding to measurement points on a wafer, FIG. 7 is a flowchart of an embodiment of rose/warp measurement according to the present invention, and FIG. 8 is an explanatory diagram of measurement points on a wafer. , FIG. 9 is a flowchart of the normal measurement procedure after calibration. 10 system, 12: vacuum chuck, 14: wafer, 16: sensor, 18: first probe, 20: second probe, 22: measurement head, 24: support, 26: reference probe, 28: reference, 3o: electron Circuit, 32: Linearizer. 34: Linearizer, 36: Linearizer, 38: Analog adder, 4o: Mixing element, 42:
A/D converter, 44: xθ2 assembly, 46:
processor, 48: data/address bus, 50: latch driver, 52: RAM. 54: PROM, 56: Central control processor, 5
8: IEEE488 bus, 60: IEEE48
8 interface.

Claims (22)

[Claims] (1) A movable holding device having a predetermined range of motion and on which a specimen to be measured is removably mounted; Holding position means that moves together with the movable holding device; A movable holding device that is remote from and within the range of motion of the movable holding device; a first sensor, the first sensor being for outputting a measurement signal of a characteristic of a specimen as the specimen is moved relative to the first sensor by the movable holding device; and; a second sensor in fixed relationship with the sensor and spaced from and within a range of motion of the movable holding device, the second sensor moving the specimen relative to the first sensor so that the second sensor is moved to the holding position by the movable holding device moving the specimen relative to the first sensor; means for providing an error signal indicative of a deviation from an ideal position of the movable holding device as the means moves relative to the second sensor; and; means for removing the error signal from the measurement signal; An error compensator for a non-measurement sample holding device having: means for providing an output signal compensated for deviations from ideal position; (2) An error compensation device for a non-measurement specimen holding device according to claim 1, wherein the movable holding device is a device movable in the x, θ, and z directions, and is equipped with a vacuum chuck that removably attaches the specimen. . (3) An error compensation device for a non-measurement sample holding device according to claim 2, wherein the holding position means is fixed to a device movable in the x, θ, and z directions. (4) An error compensation device for a non-measurement sample holding device according to claim 3, wherein the holding position means is a disk mounted on the vacuum chuck for movement in the x and θ directions. (5) An error compensation device for a non-measurement sample holding device according to claim 3, wherein the holding position means is a bar mounted on the vacuum chuck for movement in the x direction. (6) An error compensation device for a non-measurement sample holding device according to claim 1, wherein the first and second sensors are capacitive sensors. (7) The error compensating device for a non-measurement sample holding device according to claim 6, comprising a sensor support, and the first and second sensors are mounted on the sensor support. (8) a sensor; a holding device; moving the sensor and holding device relative to each other and providing a sensor measurement signal of the specimen from which one or more properties of the specimen can be determined, but the signal is non-ideal; Due to the specific holding position, the sensor is subject to errors that contaminate the sensor output, thereby reducing the accuracy of the sensor measurement signal. means; reference means representative of the ideal position of the holding device; means for measuring deviations with respect to the reference means from the ideal position of the holding device; compensating for deviations from the ideal of the measurement signal; A sample measurement apparatus for compensating for non-repeatable errors, comprising: means for providing an output signal that allows characteristics of one or more samples to be determined with high accuracy; (9) A sample measuring device for compensating for non-repeatable errors as claimed in claim 8, wherein the holding device is movable relative to the sensor. (10) A sample measuring device for compensating for non-repetitive errors according to claim 9, wherein the holding device includes a vacuum chuck movable in x, θ, and z directions. (11) Errors due to non-ideal holding positions are instantaneous and natural positional deviations of the holding device, which act to shift the specimen from its originally intended spatial position, thereby resulting in the measurement obtained by the sensor element. 11. A sample measuring device for compensating for non-repeatable errors as claimed in claim 10, wherein errors are introduced into the data. 12. A sample measuring device for compensating for non-repeatable errors as claimed in claim 11, wherein said reference means includes a disk mounted for movement with a movable holding device to define an ideal position. (13) According to claim 11, the reference means includes a bar mounted to move in the x direction together with the movable holding device and defining an ideal position, and a sensor for measuring a relative distance to the bar. A sample measuring device that compensates for non-repeatable errors as described. (14) The compensation means includes electronic circuit means for creating a compensation signal that changes as the reference means moves away from its ideal position and is measured by the sensor element to which it is associated. A sample measuring device that compensates for non-repeatable errors as described. (15) A method for compensating for contamination of measurement data caused by an abnormal position of a device that controls and moves an object, the measurement data of which is obtained by sensing the proximity of a probe, comprising: sequentially sensing different points between the object and the probe. moving the device relative to the probe to obtain measurement data by positioning it in the close range; measuring the distance of the device position from the reference position as different points of the object and the probe are successively located in the sensitive proximity range; measuring data; Remove the distance of the measured position from; a method for compensating for contamination of the measured data. (16) The measurement data contamination compensation method according to claim 15, wherein the distance between the measured positions is removed from the measurement data in real time. (17) A device for compensating for positional abnormalities of a holding device that contaminate measurement data obtained by a data probe, the holding device being moved relative to the probe, and different points of the measurement object attached to the holding device being sequentially touched by the probe. means for obtaining measurement data by positioning the object in the sensitive proximity position; measuring the distance of the holding device position from the reference position as different points of the object and the probe are successively positioned in the sensitive proximity position; A compensating device comprising: means for; and means for removing the measured position separation from the measurement data. (18) A compensation device according to claim 17, wherein the means for removing the distance of the measured position from the measurement data operates in real time. (19) Compensation according to claim 17, wherein the means for measuring the separation comprises a reference standard mounted for movement together on the holding device and a reference standard probe in fixed relationship with the data probe. Device. (20) Data of an object with an object receiver movable in the x, θ, and z directions with respect to the probe, with better accuracy than can be obtained relative to the inherent mechanical accuracy of the object receiver. An apparatus for providing measurement data, comprising: a sensor head for providing measurement data at multiple positions of an object; the object moves within the sensor head; a probe; an object receiving device movable in the x, θ, and z directions for controlling and moving multiple positions of an object within the sensor head of the probe; coupled to the probe and the object receiving device movable in the x, θ, and z directions; and controlling the operation of the object receiving device movable in the x, θ, and z directions to measure different positions of the object placed on the object receiving device within the sensor head. cyclically operating means for moving in a sequence of motion repeating the measurement cycle to obtain measurement data; coupled to said cyclically operating means for moving the x, θ, z device from an ideal holding position; means for compensating for errors in measured data due to repetitive positional deviations of the device; coupled to said cyclically operated means; The first to compensate for errors in the measurement data caused by
an apparatus comprising means and; cooperating with the means of; (21) Means for compensating for non-repetitive positional deviations of the measurement data are; mounted movably on a holding device movable in x, θ, and z directions; and a reference defining a reference position; 21. The apparatus of claim 20, further comprising: a probe for providing error data representative of the deviation of the actual position of the holding device from a reference position. (22) The device according to claim 21, wherein the means for compensating non-repetitive position deviations of the measurement data operates in real time.